Skip Navigation

Estimates of Rotavirus Disease Burden in Hong Kong: Hospital-Based Surveillance

  1. E. Anthony S. Nelson1,
  2. John S. Tam2,
  3. Joseph S. Bresee6,
  4. Kin-Hung Poon3,
  5. Chi-Hang Ng4,
  6. Kin-Sing Ip5,
  7. T. Christopher Mast7,
  8. Paul K.-S. Chan2,
  9. Umesh D. Parashar6,
  10. Tai-Fai Fok1 and
  11. Roger I. Glass6
  1. 1 Department of Paediatrics, The Chinese University of Hong Kong, Hong Kong Special Administrative Region, China
  2. 2 Department of Microbiology, The Chinese University of Hong Kong, Hong Kong Special Administrative Region, China
  3. 3 Departments of Paediatrics and Adolescent Medicine, Tuen Mun Hospital, Hong Kong Special Administrative Region, China
  4. 4 Departments of Paediatrics and Adolescent Medicine, Queen Elizabeth Hospital, Hong Kong Special Administrative Region, China
  5. 5 Departments of Paediatrics and Adolescent Medicine, Pamela Youde Nethersole Eastern Hospital, Hong Kong Special Administrative Region, China
  6. 6 Centers for Disease Control and Prevention, Atlanta, Georgia
  7. 7 Merck Research Laboratories, West Point, Pennsylvania
  1. Reprints or correspondence: Prof. E. A. S. Nelson, Dept. of Paediatrics, The Chinese University of Hong Kong, 6/F Clinical Science Bldg., Prince of Wales Hospital Shatin, Hong Kong SAR, China (tony-nelson{at}cuhk.edu.hk).

  1. Presented in part: 11th Asian Congress of Pediatrics, Bangkok, Thailand, November 2003 (abstract FO-I-1).

 
Next Section

Abstract

Background We conducted prospective, hospital-based surveillance for rotavirus disease for a 2-year period at 4 of 12 public government (Hospital Authority [HA]) hospitals in Hong Kong. It has been estimated that HA hospitals provide 90% of inpatient care in Hong Kong.

Methods Information was collected for children <5 years old who had a primary or secondary diagnosis of diarrhea or for whom a stool sample was tested for the presence of rotavirus (by enzyme immunoassay) or bacteria (by culture). Surveillance data were compared with routine discharge information from the HA's computerized Clinical Management System (CMS).

Results During a 2-year period (1 April 2001 through 31 March 2003), 7391 children were admitted to the hospital with diarrhea or developed diarrhea during their hospital stay. Of these children, 5881 (80%) had a stool sample tested for the presence of rotavirus, and 30% were positive for rotavirus (representing 24% of all diarrhea-associated admissions). CMS data underreported the total percentage of diarrhea-associated admissions (15% vs. 20%) and the percentage of diarrhea-associated admissions that were the result of rotavirus infection (13% vs. 24%). Estimated rates of hospitalization for rotavirus infection (8.8 admissions/1000 children <5 years old and 18.4 admissions/1000 children <1 year old) were 4-fold higher than our previous estimates, which were determined on the basis of CMS data alone. We estimate that the cumulative risk of hospitalization with rotavirus diarrhea by age 5 years is 1 in 24. Combined active and passive (CMS) surveillance data indicate that 4.6% of all general pediatric admissions to HA hospitals in Hong Kong were associated with rotavirus infection.

Conclusion Our study combined passive surveillance data from all Hong Kong HA hospitals with active surveillance data from 4 sentinel hospitals. The estimates of rotavirus disease burden obtained will help emphasize the effect of this important disease and create awareness of the potential for rotavirus vaccines. The surveillance model developed could also be a powerful tool for monitoring the effect of a vaccine.

Rotavirus infection is a global disease that affects almost all children and is estimated to cause approximately one-quarter of all diarrhea-related deaths (∼440,000) each year [1]. In both developed and developing countries, approximately one-third of all children admitted to the hospital with diarrhea are infected with rotavirus. Improvements in hygiene and sanitation are unlikely to affect the burden of this disease, thus emphasizing the need for a vaccine. The first licensed rotavirus vaccine, RotaShield (RRV-TV; Wyeth-Ayerst), was recommended for routine immunization in the United States in 1998 but was withdrawn in 1999 as a result of its association with intussusception [2–5]. Despite this setback, the withdrawal of RRV-TV has had a number of positive effects, including renewed interest in developing other rotavirus vaccines and the possibility that these new vaccines will be tested simultaneously in both developed and developing countries. A number of rotavirus vaccines are currently undergoing phase 2 and 3 trials, with licensure predicted within the next 2–3 years.

The World Health Organization (WHO), the Global Alliance of Vaccines and Immunizations (GAVI), and the Children's Vaccine Program (CVP) at the Program for Appropriate Health Technologies (PATH) have identified rotavirus vaccine as a priority, and significant funding has been provided to fast-track the development of a rotavirus vaccine and its introduction in developing countries [6]. To this end, the WHO has recommended the use of simple, generic protocols for surveillance, as well as the establishment of regional surveillance networks. The first such network, the Asian Rotavirus Surveillance Network (ARSN), was established in February 1999 as an initiative of the Centers for Disease Control and Prevention (CDC), WHO, the CVP at PATH, and partners from industry (GlaxoSmithKline and Merck) [7]. The initial participants in the ARSN were Hong Kong, China, Indonesia, Malaysia, Myanmar, South Korea, Taiwan, Thailand, and Vietnam.

A number of studies have assessed the epidemiological profile of rotavirus infection and other diarrheal diseases in Hong Kong [8–20]. In those studies, rotavirus infection accounted for ∼30% of cases of diarrhea among hospitalized children and ∼10% of cases of diarrhea in the community. In 1996, a computerized data collection system, the Clinical Management System (CMS), was introduced in phases to all publicly funded government (Hospital Authority [HA]) hospitals in Hong Kong, to collect uniform discharge data and other information on all patients. By use of the CMS data, we previously showed that, over the course of a 2-year period (July 1997 through June 1999), diarrhea was the primary diagnosis associated with 11.5% of all hospital admissions of children <5 years old and that 10.9% of these cases of diarrhea were due to rotavirus infection; however, the percentage of cases of diarrhea due to rotavirus infection ranged from 0% to 27.9%, depending on whether the hospital actively screened for rotavirus. This finding suggests that rotavirus infection was being underreported in the CMS data [8].

The present study was undertaken in partnership with the ARSN to obtain accurate estimates of the burden of rotavirus disease in Hong Kong and in the Asian region. The CMS has the potential to provide long-term, low-cost surveillance of rotavirus disease in Hong Kong, but this potential cannot be achieved unless the problem of underreporting of rotavirus infection through routine diagnostic coding at discharge is addressed. Thus, a secondary aim of the study was to explore the issue of the underreporting of rotavirus infection by the CMS data by comparing active surveillance data collection with passive CMS data collection. Good surveillance for rotavirus infection is seen as the key to providing information to decision makers, to ensure that future rotavirus vaccines can rapidly be made available to all children, and as the key to measuring the effect of using vaccines.

Previous SectionNext Section

Methods

The overall study design was based on the WHO Generic Protocol for conducting surveillance for rotavirus infection in developed and developing countries [21]. This core program was modified and expanded according to local needs. There are 12 HA and 10 private hospitals in Hong Kong that provide pediatric services. The active surveillance in the present study involved 4 HA hospitals, 1 in each of 4 main districts of Hong Kong: Pamela Youde Nethersole Eastern Hospital (PYNEH) on Hong Kong Island, Queen Elizabeth Hospital (QEH) in Kowloon area, Prince of Wales Hospital (PWH) in eastern New Territories, and Tuen Mun (TMH) Hospital in western New Territories. It has been estimated that 90% of inpatient care in Hong Kong takes place in HA hospitals [22], although the extent to which this estimate applies to different age groups is not known.

Clinical data

Each participating hospital has a ward logbook for recording patient identifiers and the diagnosis at admission for all children. During the study, these logbooks were viewed at least once weekly to identify all children admitted with a diagnosis of diarrhea and/or vomiting. Coordinators at each participating hospital encouraged the medical staff admitting children with diarrhea to complete a standard 2-page “Childhood Diarrhea Admission Form” that contains standardized clinical data on the child's medical history, examination findings, and details of treatments given to the child before and during hospitalization. This form was placed in the medical record and, when completed, simplified the abstraction of clinical data after discharge. Hospital admissions of children were considered to be associated with diarrhea if the children were <5 years old and had a primary or secondary diagnosis of diarrhea. The WHO Generic Protocol for conducting surveillance of rotavirus infection defined a case of rotavirus diarrhea as watery diarrhea in a child <5 years old who was admitted to a hospital in the surveillance system because of watery diarrhea and whose stool sample tested positive for rotavirus. Although acute diarrhea has been defined as an increase in stool frequency by ⩾3 unformed stools in the previous 24 h for a period of <14 days [23], we opted to leave the determination of the diagnosis at admission to admitting physicians, who would specify a diagnosis of gastroenteritis if the child had diarrhea (i.e., watery or loose stools), with or without dehydration, vomiting, or fever. Case notes for children admitted with a diagnosis of vomiting alone were reviewed to determine whether the children developed diarrhea some hours after admission. Children who did not pass any stool or who were discharged within 48 h of admission without any bowel evacuation were considered to have had a nondiarrheal hospital admission.

Laboratory data

All participating hospitals routinely sent stool samples obtained from children with diarrhea for bacterial culture, and 3 hospitals (PWH, QEH, and TMH) routinely tested for rotavirus. Rotavirus kits (DAKO) were provided to PYNEH for rotavirus testing prior to transfer of samples to PWH for genotyping. Participating hospitals entered results of both rotavirus testing and bacterial culture into a computer database. Three of the 4 hospitals (PWH, QEH, and TMH) submitted their rotavirus test results electronically on a monthly basis. PYNEH consolidated results every 6 months, or results were obtained by the research staff from the CMS. Bacterial culture results were available electronically for PWH and QEH on a monthly basis, and results from PYNEH and TMH were obtained electronically at less frequent intervals or directly from CMS. The laboratory results were then matched with the clinical surveillance data by using the patient's Hong Kong identity number or hospital number. This process identified children admitted with diarrhea for whom a stool sample was not sent for testing, and it also provided information on children with diarrhea who were missed by the ward surveillance process.

CMS data

For all 12 HA hospitals, CMS data from July 1999 to June 2003 were provided by the HA Data Centre, to enable comparison of the active surveillance data with the passive CMS data. The CMS database contains International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes that are used to define the following diarrhea-associated reasons for hospitalizations: bacterial diarrheas (codes, 001–005, excluding 003.2, and 008.0–008.5), parasitic diarrheas (codes, 006–007, excluding 006.2–006.6), rotavirus diarrhea (code, 008.61), other viral diarrheas (codes, 008.6–008.8), diarrhea of undetermined etiology (including that presumed to be infectious [codes, 009.0–009.3]), and other noninfectious diarrheas (codes, 558.9 and 787.91). The classifications were assigned according to whether these ICD-9-CM codes were the primary diagnostic code or 1 of any 14 possible discharge codes.

Incidence rates

The number of births in Hong Kong is well documented, and, although some children born in Hong Kong may reside on the Chinese mainland, it is principally children born in Hong Kong who are entitled to be admitted to HA hospitals at a nominal cost. Nonentitled children (e.g., children of visitors to Hong Kong) are more likely to be admitted to private hospitals. A smaller number of children not born in Hong Kong are entitled to admission to HA hospitals if one parent is a Hong Kong resident. This information was recorded by the Integrated Patient Administration System and was linked to the CMS data. The incidence rates for rotavirus-related diarrhea hospitalizations were calculated using the number of admissions to HA hospitals and the number of births in Hong Kong. Although this estimate is limited by the absence of data from the private hospital system, previous studies have estimated that ∼90% of inpatient care in Hong Kong is provided within the HA system [22].

Data analysis

Data were entered using Epi Info (version 6.04d; CDC). Core surveillance data, with all patient identifiers removed, were sent to the CDC on a monthly basis for further grouped analyses with data from the other ARSN participants.

Ethics

The study was approved by the Clinical Research Ethics Committee of the Chinese University of Hong Kong, as well as the corresponding ethics committees of individual participating hospitals.

Previous SectionNext Section

Results

Although surveillance started on 1 December 2000, the analysis focused on data collected during the 24-month study period (1 April 2001 through 31 March 2003). The period from April through March each year (rather than January through December) was selected for analysis, because the rotavirus infection season in Hong Kong peaks from November through February. During this 24-month period, there were 8769 pediatric diarrhea-related admissions, and 7779 (89%) of the patients admitted were <5 years old. Of the children <5 years old, 388 had vomiting alone without diarrhea or did not pass a stool after admission, leaving 7391 patients eligible for inclusion. These patients were distributed between the 4 participating hospitals as follows: PWH, 1708 (23%); TMH, 2571 (35%); QEH, 1782 (24%); and PYNEH, 1330 (18%). The childhood diarrhea form was completed for 37% (2716/7391) of patients admitted with diarrhea. Four hundred seventy-one patients participated in a substudy that collected data on the economic burden of rotavirus. Seventy percent (5130/7294) of patients had a primary diagnosis at admission that included gastroenteritis, diarrhea, gastritis, or vomiting. Eleven percent of children (826/7264) had previously been admitted within the past 3 weeks, and, for 311 children (38%), this admission was due to diarrhea.

Admissions to HA hospitals are assigned various entitlement codes, depending on whether the child or the parents are entitled to receive subsidized treatment. As stated above, nonentitled children are usually those born outside Hong Kong, although children born outside Hong Kong are entitled if either parent has a Hong Kong identity card. The detailed breakdown of the entitlement codes was as follows: 5820 entitled persons (82%), 139 nonentitled persons whose parent was entitled (2%), 12 dependents of HA staff (0.2%), 70 government servants or dependents (0.9%), 71 persons receiving public assistance (0.9%), 45 nonentitled foreign persons or 2-way permit holders from mainland China (0.6%), and 985 unidentified persons (14%).

Overall, during the study period, 80% (5881/7391) of patients had a stool sample sent for rotavirus testing, and rotavirus antigen was identified in 30% (1758/5879) of samples tested, which accounted for 24% (1758/7391) of all admissions. The case notes for 75% (5412/7263) of the patients indicated that a stool sample had been sent for rotavirus testing, but, for 5.2% (280/5412) of these patients, no such record could be found in the laboratory. Conversely, laboratory records indicated that rotavirus testing was performed for 642 (35%) of 1842 stool samples for which 25% of case notes did not state that a sample had been sent for testing. Stool samples from 87% (6451/7391) of the patients were sent for bacterial culture, and a pathogen was identified in 24% (1534/6451) of samples tested, which accounted for 21% (1534/7391) of all admissions. Of the bacterial culture-positive samples, salmonella was identified in 44%, campylobacter was identified in 11%, shigella was identified in 1%, “other” pathogens were identified in 33%, and >1 pathogen was identified in 11%. Twelve percent (204/1722) of the rotavirus-positive stool samples also were positive for a bacterial pathogen. A record indicating that a stool sample had been sent for bacterial culture was found in 78% (5671/7263) of case notes. The proportion of case notes with a record indicating that a stool sample had been sent to the laboratory and had been received by the laboratory was similar among patients who had a diagnosis related to diarrhea at admission and those who did not (table 1). However the proportion of positive results, for both rotavirus testing and bacterial culture, was higher among patients with a diagnosis of gastroenteritis at admission. As in previous studies in Hong Kong [24], there was a consistent excess of male patients among all patients with diarrhea-related admissions (59%), among rotavirus-positive patients (55%), and among rotavirus-negative patients (60%).

View this table:
Table 1

Results of rotavirus testing and bacterial culture for stool samples obtained from children admitted to surveillance hospitals in Hong Kong, by primary diagnosis at the time of admission.

We examined the cumulative percentage of admissions, by age (in months), for those children who were rotavirus positive and rotavirus negative (figure 1). Children <2 years old accounted for 71% of all children admitted for diarrhea, 68% of patients with rotavirus-positive stool samples, 84% of patients with positive bacterial culture results, and 69% of patients whose results for both rotavirus testing and bacterial culture were negative.

Figure 1
View larger version:
Figure 1

Cumulative percentage of rotavirus (RV)-positive (RV+) children admitted to surveillance hospitals (solid line) and cumulative no. of RV tests performed (diamonds), by age of the children.

The seasonality of admissions for patients with rotavirus-positive and -negative diarrhea was examined over the course of the 24-month period from 1 April 2001 through 31 March 2003 (figure 2). Rotavirus infection had a clear winter seasonal peak from January through March and was infrequently identified in the summer season. In contrast, cases of rotavirus-negative diarrhea peaked in the summer but were present year round.

Figure 2
View larger version:
Figure 2

Seasonality of no. of diarrhea-related hospital admissions (solid line) and percentage of rotavirus (RV)-positive (RV+) patients (squares), from 1 April 2001 through 31 March 2003.

Three children died during the study period. One child was admitted to the hospital and died on the day of admission with a diagnosis of gastroenteritis and a 1-day history of vomiting and diarrhea; no stool sample was sent for testing. Another child died a month after admission with a primary diagnosis of neuroblastoma, although the child had diarrhea 3 days after admission; the child's stool sample was negative for rotavirus and bacteria. Clinical details for the third child were not available, and no stool samples were sent for testing.

Surveillance data were compared with CMS data for the 24-month period from 1 April 2001 through 31 March 2003. In the 4 surveillance hospitals, 13% of all admissions had a primary diarrhea code, 15% had either a primary or secondary code indicating diarrhea, 1.7% of all admissions had a primary code for rotavirus (ICD-9-CM code 008.61), and 2.0% had any diagnostic code indicating rotavirus (table 2). However, the active surveillance data identified rotavirus in the stool samples of 4.6% (1758/37,829) of all general pediatric admissions to the 4 surveillance hospitals. On the basis of these active and passive surveillance data for this 24-month period, incidence rates of rotavirus-associated admissions were calculated to be 8.1–8.8 admissions/1000 children <5 years old and 15.4–18.4 admissions/1000 children <1 year old (table 3). On the basis of the passive CMS surveillance data alone, the corresponding incidence rates of rotavirus disease (any of 14 diagnostic codes) were 2.9–3.0 cases/1000 children <5 years old and 5.4–5.5 cases/1000 children <1 year old (table 3). When only the primary diagnostic code was used, the corresponding incidence rates were 2.4–2.5 cases/1000 children <5 years old and 4.2–4.4 cases/1000 children <1 year old.

View this table:
Table 2

Comparison of passive Clinical Management System (CMS) and active surveillance data from 4 Hong Kong government hospitals participating in a surveillance study (n = 37,829 admissions) and 8 hospitals not participating in the surveillance study (n = 57,382 admissions) from 1 April 2001 through 31 March 2003.

View this table:
Table 3

Incidence rate (IR) of rotavirus infection per 1000 children aged 1–59 months admitted to Hong Kong government (Hospital Authority [HA]) hospitals from 1 April 2001 through 31 March 2003.

Previous SectionNext Section

Discussion

The present study has documented the importance of diarrheal diseases and rotavirus infection as leading causes of hospital admission in Hong Kong and has confirmed the findings of previous reports [11]. Diarrhea was extremely common, with 20% of all general pediatric admissions to the 4 surveillance hospitals being associated with diarrhea; 80% of the patients admitted had a stool sample sent for rotavirus testing. Of all the patients with diarrhea, excluding those who had no bowel movement after admission but including those admitted for other reasons who had stools samples sent for rotavirus testing and bacterial culture, 24% were positive for rotavirus, and 30% of stool samples that actually were tested for rotavirus were positive. Previous studies in Hong Kong undertaken since the mid-1980s have shown that 26%–35% of inpatient stool samples tested were positive for rotavirus [9].

These results confirm our previous concerns about the underreporting of rotavirus disease by the CMS [8]. In addition, the data show that all causes of diarrhea are underreported by routine discharge diagnostic coding (table 2). During the period from 1 April 2001 through 31 March 2003, the CMS data significantly underreported both the total number of diarrhea-associated admissions (15% vs. 20%) and the percentage of diarrhea-associated admissions for which a stool sample was found to be positive for rotavirus (13% vs. 24%). The underestimation of the percentage of diarrhea-associated admissions for which a stool sample was found to be positive for rotavirus is likely to have been even greater had all diarrheal stool samples been tested for rotavirus (i.e., 30% of samples tested were positive). If 30% of all untested stool samples were also positive for rotavirus, then 5.9% of all general pediatric admissions would have been associated with rotavirus infection. Conversely, it is possible that a proportion of children with rotavirus identified in a stool sample were admitted for reasons other than rotavirus diarrhea and that infection with the pathogen was not clinically significant. It is interesting to note that the proportion of admissions coded as being associated with rotavirus infection was similar in the 4 surveillance and 8 nonsurveillance hospitals (table 2). This result suggests that active surveillance itself did not have a major effect on the coding process.

By combining the active surveillance data with the CMS discharge data for all HA hospitals, it has been possible to improve our estimate of the incidence of rotavirus infection (table 3), which is significantly higher than that previously reported when only CMS discharge diagnostic coding was used [8]. The incidence estimates in table 3 do not take into account the number of children admitted to private hospitals in Hong Kong (estimated to be ∼10% of all admissions). For the calculation of incidence data, we used the number of births in Hong Kong as the denominator. The eligibility criteria showed that the percentage of admissions of nonentitled persons was small, although 14% of admitted patients were classified as unidentified. The latter group may include entitled children admitted without evidence of their or their parents' status as residents of Hong Kong, or they may not be nonentitled. It is recognized that a significant number of children born in Hong Kong travel back and forth between mainland China and Hong Kong and may live for extended periods on the mainland with their parents or other relatives. However, there have been no reported studies that have determined the extent of this practice. Therefore, it is likely that a proportion of children born in Hong Kong will seek medical care for diarrhea while staying on the mainland and will not be included in these incidence calculations. Countering this effect is the fact that this analysis did not determine the proportion of case patients who were readmitted for the same diarrheal episode, a factor that would result in the overestimation of the true incidence of diarrheal illness. Also not considered in these incidence calculations is the fact that, during the past 6 years, ∼7000 children <5 years old have immigrated to Hong Kong from the Chinese mainland each year. Acknowledging these limitations, our rates of hospitalization for rotavirus infection in Hong Kong (8.1–8.8 admissions/1000 children <5 years old and 15.4–18.4 admissions/1000 children <1 year old) are ∼4 times higher than our previous estimates (1.7–2.8 admissions/1000 children <5 years old and 4.4–6.1 admissions/1000 children <1 year old), which were based only on the CMS data. These annual rates are also higher than those reported in other industrialized countries for children <5 years old (e.g., 3.7 admissions/1000 children in Sweden [25], 5 admissions/1000 children in England and Wales, and 4.8 admissions/1000 children in Denmark). In the United States, 1 in 78 children is hospitalized for rotavirus diarrhea [26]. Taking into account the likely underreporting of rotavirus infection in the CMS, we considered that the cumulative risk of hospitalization for rotavirus diarrhea by age 5 years could be as high as 1 in 28 children [8]. Our current data now confirm this estimate, giving a risk of 1 in 24 children (1 in 23 children in 2001 and 1 in 25 children in 2002) (table 3). If new immigrant children are included in the calculation, the risk changes to 1 in 27 children (1 in 26 children in 2001 and 1 in 28 children in 2002).

It is also possible that the incidence estimates for rotavirus infection would have been higher if stool samples from all children with diarrhea had been tested. There were 3 main reasons why stool samples were not obtained for laboratory analysis during the study period: (1) the child's diarrhea resolved quickly, and the next stool was formed, rather than loose; (2) the child's general condition improved rapidly (with or without continued diarrhea), and the child was allowed to return home pending review the following day; or (3) the child's parents chose to seek medical attention elsewhere shortly after admission, and the child was discharged before a stool sample was obtained. Although the active surveillance process included steps to ensure that stool samples from as many children as possible were sent for testing, it proved to be difficult to increase the rate of testing for rotavirus. It was noted, however, that the percentage of stool samples sent for bacterial culture was consistently greater than the percentage sent for rotavirus testing, possibly reflecting the fact that bacterial culture of stool samples has long been routine in many hospitals, whereas rotavirus testing has not. Bacterial culture was performed on stool samples obtained from 87% of patients with diarrhea-associated admissions, and 24% of stool samples tested had positive results. This percentage was less than that found in our previous study [11], which showed that, for approximately one-third of the admissions, a pathogen was isolated, most of which were Salmonella species (which accounted for 25% of all admissions). Twelve percent of the rotavirus-positive case patients were coinfected with a bacterial pathogen; some of those cases of coinfection might have been the result of nosocomial infection. Thirty percent of the children with diarrhea-associated admissions were identified as a result of stools samples being sent to the laboratory. Eleven percent of children had recently been admitted to the hospital, and, for 38% of them, the previous admission also was related to diarrhea.

Ideally, CMS could provide a low-cost surveillance tool that is specific but not completely sensitive. However, to achieve this objective, more attention to diagnostic coding at discharge is required. The ICD-9-CM codes entered into the CMS are selected by the responsible medical officer at the time of a patient's discharge from the hospital. The CMS allows the physician to type in the ICD-9-CM code directly, if known, or to use a keyword search. The ICD-9-CM codes entered may thus be influenced by a number of factors, including the availability of the patient's laboratory results at the time of discharge. It might be possible to modify the system to prompt the physician for certain conditions that are part of surveillance systems. Alternatively, if a nonspecific code is entered (e.g., ICD-9-CM code 009.3), the physician could be prompted with a list of more-specific codes. Linkage of the laboratory data with the discharge codes could provide auditing and feedback to the clinical staff, as well as a reliable estimate of the total number of admissions for which particular pathogens are isolated. The use of such data for estimating the incidence of a range of infective conditions would depend on the extent to which samples are routinely sent to the laboratory. At present, many clinical departments will not routinely request such investigations, because the results of laboratory testing do not usually influence treatment of the patient, particularly when cost savings are being sought. A possible solution to address this issue would be to identify sentinel hospitals that would be required to routinely test all stool samples. Our model of combining sentinel active surveillance with universal passive surveillance warrants further study to determine the minimum level of active surveillance required to produce reliable estimates of disease burden. Such a surveillance scheme is more likely to be sustainable in the long term.

The present study has confirmed that rotavirus infection accounts for approximately one-third of all diarrhea-associated hospital admissions in Hong Kong and that 20% of all general pediatric admissions are related to diarrhea (either as a primary or secondary diagnosis). The cumulative risk of hospitalization associated with rotavirus diarrhea by 5 years of age is estimated to be 1 in 24 children. A future safe and effective rotavirus vaccine could have a major effect on this situation. Advocacy by opinion leaders and local experts using good surveillance data will be essential to create awareness of and demand for rotavirus vaccine, and good surveillance data would be a powerful tool for monitoring the effect of a vaccine.

Previous SectionNext Section

Acknowledgments

We thank the Hospital Authority Head Office, for providing data from the Clinical Management System; Fung Hong, for providing advice; Agnes C. K. Cheng, Tracey C. S. Tsen, and Joyce Sau Wa Lee, for helping with the initial piloting of study instruments and data collection; Que Tak Lun, Yip Kam Tong, and Chan Yin Leung Benedict (Tuen Mun Hospital); Frankie Ng and Donny Lyon (Prince of Wales Hospital); Dominic N. C. Tsang and Cho Chiu Tiu (Department of Pathology, Queen Elizabeth Hospital); Raymond Yung (Department of Pathology, Pamela Youde Nethersole Eastern Hospital), for providing help with laboratory data collection from the 4 surveillance hospitals; and Claudia Chesley, for giving editorial comments.

Previous SectionNext Section

Footnotes

  • Financial support: Hong Kong Research Grants Council; World Health Organization Department of Vaccines and Biologicals; Merck & Co., Inc.

  • Potential conflicts of interest: T.C.M. holds equity interest in Merck. E.A.S.N. has received funding and support from Merck for rotavirus surveillance studies, is currently principal investigator of a phase 3 rotavirus vaccine study funded by GlaxoSmithKline, and has received lecture fees and travel support from GlaxoSmithKline. All other authors: no potential conflicts reported.

Previous Section
 

References

  1. ↵
    1. Parashar UD,
    2. Hummelman EG,
    3. Bresee JS,
    4. Miller MA,
    5. Glass RI
    . Global illness and deaths caused by rotavirus disease in children. Emerg Infect Dis 2003;9:565-72.
    MedlineWeb of Science
  2. ↵
    Centers for Disease Control and Prevention. Withdrawal of rotavirus vaccine recommendation. MMWR Morb Mortal Wkly Rep 1999;48:1007.
    Medline
  3. Centers for Disease Control and Prevention. Intussusception among recipients of rotavirus vaccine—United States, 1998–1999. MMWR Morb Mortal Wkly Rep 1999;48:577-81.
    Medline
    1. Murphy TV,
    2. Gargiullo PM,
    3. Massoudi MS,
    4. et al
    . Intussusception among infants given an oral rotavirus vaccine. N Engl J Med 2001;344:564-72.
    CrossRefMedlineWeb of Science
  5. ↵
    1. Kramarz P,
    2. France EK,
    3. Destefano F,
    4. et al
    . Population-based study of rotavirus vaccination and intussusception. Pediatr Infect Dis J 2001;20:410-6.
    CrossRefMedlineWeb of Science
  6. ↵
    1. McCarthy M
    . Project seeks to “fast track” rotavirus vaccine. Lancet 2003;361:582.
    MedlineWeb of Science
  7. ↵
    1. Bresee JS,
    2. Fang ZY,
    3. Wang B,
    4. et al
    . First report from the Asian Rotavirus Surveillance Network. Emerg Infect Dis 2004;10:988-95.
    MedlineWeb of Science
  8. ↵
    1. Nelson EA,
    2. Tam JS,
    3. Glass RI,
    4. Parashar UD,
    5. Fok TF
    . Incidence of rotavirus diarrhea and intussusception in Hong Kong using standardized hospital discharge data. Pediatr Infect Dis J 2002;21:701-3.
    CrossRefMedlineWeb of Science
  9. ↵
    1. Nelson EA,
    2. Tam JS,
    3. Yu LM,
    4. Glass RI,
    5. Parashar UD,
    6. Fok TF
    . Surveillance of childhood diarrhoeal disease in Hong Kong, using standardized hospital discharge data. Epidemiol Infect 2004;132:619-26.
    Medline
    1. Chan PK,
    2. Tam JS,
    3. Nelson EA,
    4. et al
    . Rotavirus infection in Hong Kong: epidemiology and estimates of disease burden. Epidemiol Infect 1998;120:321-5.
    CrossRefMedline
  11. ↵
    1. Biswas R,
    2. Lyon DJ,
    3. Nelson EA,
    4. Lau D,
    5. Lewindon PJ
    . Aetiology of acute diarrhoea in hospitalized children in Hong Kong. Trop Med Int Health 1996;1:679-83.
    MedlineWeb of Science
    1. Chan PK,
    2. Lim WL,
    3. Wong MF,
    4. Kwok CSY
    . Viral gastroenteritis in Hong Kong. J Hong Kong Med Assoc 1994;46:195-9.
    1. Ling JM,
    2. Cheng AF
    . Infectious diarrhoea in Hong Kong. J Trop Med Hyg 1993;96:107-12.
    MedlineWeb of Science
    1. Tam JS
    . Infantile diarrhoea: a prospective study. Hong Kong: University of Hong Kong; 1990. p. 1-192. [PhD thesis].
    1. Zheng BJ,
    2. Lo SK,
    3. Tam JS,
    4. et al
    . Prospective study of community-acquired rotavirus infection. J Clin Microbiol 1989;27:2083-90.
    Abstract/FREE Full Text
    1. Lam BC,
    2. Tam JS,
    3. Ng MH,
    4. Yeung CY
    . Nosocomial gastroenteritis in paediatric patients. J Hosp Infect 1989;14:351-5.
    CrossRefMedlineWeb of Science
    1. Chan RC,
    2. Tam JS,
    3. Fok TF,
    4. French GL
    . RNA-electrophoresis as a typing method for nosocomial rotavirus infection in a special-care baby unit. J Hosp Infect 1989;13:367-75.
    CrossRefMedlineWeb of Science
    1. Tam JS,
    2. Zheng BJ,
    3. Yan YK,
    4. et al
    . Occurrence of rotaviruses in Guangzhou and Hong Kong. J Infect Dis 1988;157:357-63.
    Abstract/FREE Full Text
    1. Yam WC,
    2. Lung ML,
    3. Yeung CY,
    4. Tam JS,
    5. Ng MH
    . Escherichia coli associated with childhood diarrheas. J Clin Microbiol 1987;25:2145-9.
    Abstract/FREE Full Text
  20. ↵
    1. Tam JS,
    2. Kum WW,
    3. Lam B,
    4. Yeung CY,
    5. Ng MH
    . Molecular epidemiology of human rotavirus infection in children in Hong Kong. J Clin Microbiol 1986;23:660-4.
    Abstract/FREE Full Text
  21. ↵
    World Health Organization (WHO). Generic protocol for (i) hospital-based surveillance to estimate the burden of rotavirus gastroenteritis in children and (ii) a community-based survey on utilization of health care services for gastroenteritis in children. Field test version (WHO/V&B/02.15). Geneva: WHO; 2002.
  22. ↵
    1. Chiu SS,
    2. Lau YL,
    3. Chan KH,
    4. Wong WH,
    5. Peiris JS
    . Influenza-related hospitalizations among children in Hong Kong. New Engl J Med 2002;347:2097-103.
    CrossRefMedlineWeb of Science
  23. ↵
    World Health Organization. The epidemiology and etiology of diarrhoea. Geneva: World Health Organization; [Accessed 7 June 2005]. Available at: http://www.who.int/child-adolescent-health/New_Publications/CHILD_HEALTH/Meded/1med.htm.
  24. ↵
    1. Nelson EA,
    2. Mok TC,
    3. Yu LM
    . Retrospective comparison of management of gastro-enteritis in hospitalised children. Ann Trop Paediatr 2002;22:165-71.
    CrossRefMedlineWeb of Science
  25. ↵
    1. Johansen K,
    2. Bennet R,
    3. Bondesson K,
    4. et al
    . Incidence and estimates of the disease burden of rotavirus in Sweden. Acta Paediatr Suppl 1999;88:20-3.
    CrossRefMedline
  26. ↵
    Rotavirus vaccine for the prevention of rotavirus gastroenteritis among children: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Recomm Rep 1999;48(RR-2):1-20.
| Table of Contents

This Article

  1. J Infect Dis. (2005) 192 (Supplement 1): S71-S79. doi: 10.1086/431492
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 205 (5)
  1. Journal of Infectious Diseases
  1. Alert me to new issues

Most